从 Dataweave 中的列中获取最大值

Question

输入：

number,fruit,colour,join_date,avginqty,avgoutqty,resolution
282,apple,red,today,3 qty,2 qty,640*320
282,apple,red,today,3 qty,,200*200
282,apple,red,today,1 qty,,640*320
282,banana,yellow,yesterday,,,
282,banana,yellow,yesterday,1 qty,,200*200
/283,banana,yellow,yesterday,1 qty,,200*200

输出 Json：

number,fruit,colour,join_date,avginqty,avgoutqty, resolution
282,apple,red,today,2 qty,2 qty, 640*320
282,banana,yellow,yesterday,1 qty,,200*200

我想得到以下内容：

• 获取最常见的分辨率列数据
• 并删除数字列中包含“/”的行

数量：

(payload groupBy (key,index) ->(key.number ++ key.join_date)) mapObject {
   a: {
       number: $[0].number,
       fruit: $[0].fruit,
       colour: $[0].colour,
       join_date:$[0].join_date,
       avginqty: if((sizeOf($.avginqty filter $ != "")) != 0) ($.avginqty sumBy (if($ == "") 0 else $))/(sizeOf($.avginqty filter $ != "")) ++ " qty" else "", 
       avgoutqty: if((sizeOf($.avgoutqty filter $ != "")) != 0) ($.avgoutqty sumBy (if($ == "") 0 else $))/(sizeOf($.avgoutqty filter $ != "")) ++ " qty" else ""
   }
} pluck $

Answer 1

下面的脚本可以满足您的要求，但不会比较分辨率。

输入

number,fruit,colour,join_date,avginqty,avgoutqty,resolution
282,apple,red,today,3 qty,2 qty,200*200
282,apple,red,today,3 qty,,640*320
282,apple,red,today,4 qty,,641*320
282,apple,red,today,1 qty,,640*320
282,banana,yellow,yesterday,,,
282,banana,yellow,yesterday,2 qty,,200*200
282,banana,yellow,yesterday,1 qty,,200*200
/283,banana,yellow,yesterday, qty,,200*200

脚本

%dw 2.0
output application/csv
import * from dw::core::Arrays
var countOfResolution = (payload filter (!($.number contains "/")) groupBy ($.number ++ $.fruit ++ $.resolution) mapObject {
    (($.number[0]++"_"++$.fruit[0]) :
      { 
          resolution: $.resolution[0],
          occurence:$.resolution reduce (( curVal, acc= 0 ) -> 
    acc + 1 )
      }
    ) 
} orderBy -$.occurence)

---
payload filter (!($.number contains "/")) groupBy ($.number ++ $.fruit ++ $.resolution) mapObject {

        (a: {
            number: $[0].number,
            fruit: $[0].fruit,
            colour: $[0].colour,
            join_date:$[0].join_date,
            avginqty:  (if((sizeOf($.avginqty filter $ != "")) != 0 ) ($.avginqty sumBy (if(($ splitBy " ")[0] == "") 0 else ($ splitBy " ")[0]))/(sizeOf($.avginqty filter $ != "")) ++ " qty" else ""), 
            avgoutqty: if((sizeOf($.avgoutqty filter $ != "")) != 0) ($.avgoutqty sumBy (if(($ splitBy " ")[0] == "") 0 else ($ splitBy " ")[0]))/(sizeOf($.avgoutqty filter $ != "")) ++ " qty" else "",
            resolution: $.resolution[0] 
        }) if((countOfResolution.*"$($[0].number ++"_"++ $[0].fruit)" filter $.resolution != "")[0].resolution == $.resolution[0] )
    } 
    pluck $

输出

number,fruit,colour,join_date,avginqty,avgoutqty,resolution
282,apple,red,today,2 qty,,640*320
282,banana,yellow,yesterday,1.5 qty,,200*200

脚本分解

1. 创建一个对象来保存每个 number_fruit 组合的分辨率及其出现次数。
2. 按降序对结果对象进行排序，以获得出现次数最多的分辨率。
3. 在主脚本中，如果当前 分辨率等于最大对象的输出 为给定的 number_fruit 计算发生的分辨率 组合。

从这里开始尝试并根据您的需要对其进行增强。如果您对此有所不同，请务必将其作为单独的问题提出。

Answer 2

您只需在脚本中添加一个过滤器即可删除斜线：​​payload filter (!($.number contains "/")

并不是说我已经从输入量中删除了" qty" 字符串，以便可以对它们求和。在同一列中混合数字和字符串对我来说没有意义。

完整脚本：

%dw 2.0
import * from dw::core::Arrays
output application/csv
---
(payload filter (!($.number contains "/"))
    groupBy (key,index) ->(key.number ++ key.join_date))

    mapObject {
        a: {
            number: $[0].number,
            fruit: $[0].fruit,
            colour: $[0].colour,
            join_date:$[0].join_date,
            avginqty: if((sizeOf($.avginqty filter $ != "")) != 0) ($.avginqty sumBy (if($ == "") 0 else $))/(sizeOf($.avginqty filter $ != "")) ++ " qty" else "", 
            avgoutqty: if((sizeOf($.avgoutqty filter $ != "")) != 0) ($.avgoutqty sumBy (if($ == "") 0 else $))/(sizeOf($.avgoutqty filter $ != "")) ++ " qty" else ""
        }
    } 
    pluck $